Here's what's new in our latest update:
​

AI Assistant Updates

Web Search

The AI assistant can now search the web in real time to answer questions about current information — such as the latest software documentation, recent research, or up-to-date product details — going beyond what is stored in its knowledge base. This capability is available across all main-app assistant agents, including the supervisor, setup, data analytics, and recommender agents.
​

Latest Models

The AI assistant now uses latest Opus 4.7 and GPT 5.5 models for various tasks on the platform

Here's what's new in our latest update:
​

New Colormap

You can select Fast from the colormap picker in the visualization UI, alongside existing options like Rainbow, Cool to Warm, and Plasma.
​

Deduplicate Pipeline Job Creation

An optional `request_id` parameter is now available. This can be used as an idempotency key to deduplicate pipeline/job creation across multiple logically identical requests.

Here's what's new in our latest update:
​

Automatic Streamline Seed Placement

The Python SDK now includes AutoStreamlines, a new filter that automatically determines optimal seed placement for streamline visualizations using a data-driven approach. This removes the need to manually configure seed positions, making it easier to generate accurate flow visualizations with less setup.

Web Search in the Notebook AI Assistant

The Notebook Assistant is able to do web searches to fetch the latest information about coding tasks and others.

Here's what's new in our latest update

Meshing Performance

We’ve significantly accelerated meshing performance. Across benchmarks, mesh time improved by 2.5x to 6x (10M–100M CVs), cutting total simulation time — so you can generate high‑fidelity results faster without changing mesh quality.

Pipelines

Pipelines is a new Early Access capability for automating large batches of simulation work on Luminary. Define a reusable, parameterized workflow as a set of dependent stages (e.g., ReadGeometry → Mesh → Simulate), then run hundreds to thousands of variants by invoking it with a PipelineArgs table—each row becomes a run.

Pipelines integrates with Named Variables so each run can apply a different Named Variable Set to generate parametric geometry while the UI provides dashboards and detailed views to manage execution and results.

To help you get started, a new Pipelines Tutorial sample notebook has been added. It walks through a complete end-to-end Pipelines workflow on Luminary using a parametric wing-body geometry, and can be opened and cloned directly from the UI.

AI Assistant Upgrades

The AI assistant in the can now browse and read your notebooks and uploaded files on demand. When you ask the assistant a file you've uploaded, it can look up the contents directly.
​
The AI Assistant now shows its thinking process while generating responses.

Fullscreen Support

The 3D visualization canvas now supports a fullscreen mode, maximizing your screen real estate for simulation analysis. Click the new fullscreen button in the canvas toolbar to expand the 3D view while keeping the toolbar, results, and geometry panels accessible. In fullscreen mode, you can still step through simulation iterations, access the AI assistant, and control the camera — without needing to exit fullscreen.

Bug Fixes

Here's what's new in our latest update

Unit Systems

When creating a project, you can now specify your preferred unit system:

• SI Units (default) - Meters, Kilograms, Kelvin, and standard SI units
• US Units - Feet, Pounds, Fahrenheit, PSI, and other US customary units
• Custom Units - Define your own units for each quantity individually

Note: Unit systems are set at project creation time and cannot be changed later. This ensures consistency and accuracy throughout your entire simulation workflow.

Automatic Unit Conversion

• All geometry features are automatically converted to your project's units when imported
• Visualization data and mesh coordinates scale appropriately
• Simulation results display in your chosen units throughout the platform

LMA Active Regions UI

Define geometric zones (boxes, cylinders, spheres) to control adaptive mesh refinement with fine-grained error weighting and distance parameters.

Key Features:

• Create and manage active regions directly in the mesh tree
• Per-region controls: visibility toggle, rename, duplicate, delete
• Configure error weight and max distance for mesh adaptation falloff
• Real-time 3D visualization in LCVis
• Visibility states persist per-project

Available when using LMA (Automatic Mesh Refinement) as your meshing method.

Conversation Compaction

The Assistant now automatically compacts long conversation histories to stay within the model's context window. When compacting runs, a "Compacting conversation" status indicator appears in the chat.

Here's what's new in our latest update:
​

Updated UI for Named Variables

Ability to Invert Y-values in the Output Charts

Support for Screenshot Upload to the Assistant

You can now upload screenshots of the app to assistant for better context.

Intersection Curves Across Multiple Slices

Intersection curves now automatically merge and connect across multiple surfaces, ensuring smooth, continuous polylines for accurate data extraction.

Here's what's new in our latest update:
​

Transform Support in Results View

We’ve introduced a new transformation feature that enables users to mirror or pattern simulation results (such as half-plane or half-car models) to visualize the full domain without additional computational cost. By applying these transforms to surfaces and filters in the results tab, users can easily generate high-fidelity, complete visualizations for symmetric or rotationally symmetric simulations.

Cancel/Delete Individual Runs in DoE

To improve control over high-volume workflows, we have decoupled individual runs within Design of Experiments (DOE) sets, allowing users to manage them independently. Users can now cancel or delete specific runs within a DOE without affecting the entire batch.

Here's what's new in our latest update:
​

Shock Capturing Options

The FDS (Flux Difference Splitting) convective scheme now includes a `shock_capturing` option in `SpatialDiscretizationFluid`. When enabled, the solver automatically activates a more robust convective scheme in regions where shocks are detected or where absolute pressure drops near a user-defined realizability limit.

SDK Code Generation for Visualization

You can now generate SDK code for visualization

​

Here's what's new in our latest update:
​

Real Gas Support

Luminary Cloud now supports Real Gas modeling, enabling simulations with equations of state that account for non-ideal gas behavior. This brings improved accuracy for high-pressure applications, cryogenic flows, supercritical fluids, and phase-change scenarios

To get started, select "Real Gas" as your fluid model in your simulation setup, choose your preferred equation of state, and configure fluid properties

Download Design of Experiment Table

You can now download DoE tables as a comma separated value (csv).
​

Here's what's new in our latest update:
​

Force Distribution Plots

You can plot force distribution plots to visualize the spanwise lift distribution, accumulated drag, or general force generated by selected surfaces.

UI Improvements

You can now use keyboard shortcuts for axis-aligned camera views. You can now press **X**, **Y**, or **Z** to quickly snap the camera to the corresponding axis view. Pressing the same key again toggles between the positive and negative directions (e.g., press X for +X, press X again for −X). These shortcuts are also displayed in the camera views menu for easy reference.

Bug Fixes

Code generation bug fixes related to geometry export.